PRENATAL DIAGNOSIS, VOL.
12,93-10 1 ( 1992)
AMNIOTIC FLUID LEVELS OF HUMAN CHORIONIC GONADOTROPIN AND ITS ALPHA AND BETA SUBUNITS IN SECOND-TRIMESTER CHROMOSOMALLY ABNORMAL PREGNANCIES c. WOLF*, FRANCIS w. BYRN*,
THOMAS s. MCCONNELL~AND M. B. KHAZAELI~ *Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology and ?Department of Pathology, University of New Mexico Schoolof Medicine, New Mexico, U.S.A.;and SDeparrment of Medicine. University of Alabama at Birmingham, Alabama. V.S.A.
GORDON
SUMMARY Amniotic fluid from 135 pregnancies was assayed for human chorionic gonadotropin (hCG) and its free alpha (ahCG) and free beta (bhCG) subunits. Forty-six chromosomally abnormal pregnancies between 14 and 20 weeks’ gestation were matched with 89 chromosomally normal samples. Compared with controls, trisomy 21 pregnancies exhibited significantly elevated levels of all three peptides, whereas trisomy 18 gestations gave rise only to significant elevation of ahCG. Female fetuses in both the trisomy 21 and trisomy 18 pregnancies provided significantly elevated levels of hCG and bhCG compared to their male counterparts. On converting the values to multiples of the median, it was determined that 6 of 7 trisomy 18 samples had abnormally elevated alpha/beta ratios, as did 6 of 21 Down’s syndrome pregnancies. Further, 11 of 21 trisomy 21 gestations had abnormal amniotic fluid hCG levels. Using only ahCG, bhCG and their ratio, a 61 per cent sensitivity was found for these trisomies, with a 96 per cent specificity. KEY WORDS
Amniotic fluid hCG and subunits
INTRODUCTION In an effort to prenatally diagnose chromosomally anomalous pregnancies, various chemical and hormonaI parameters have recently come under scrutiny. These include alpha-fetoprotein (AFP) (Merkatz et al., 1984), unconjugated oestriol (UE3) (Canick et al., 1988), human chorionic gonadotropin (hCG) and its alpha subunit (Bogart et al., 1987; Petrocik et al., 1989) and beta subunit (Berkowitz et al., 1989), cancer antigen 125 (CA 125) (Check et al., 1990), Schwangershaft protein 1 (SPI) (Bartels and Lindemann, 1988), and combinations thereof (Wald et al., 1988; O’Brien et al., 1990; Petrocik et al., 1990). Evaluation of the association between low material serum AFP levels and fetal chromosome abnormalities suggests that approximately 30 per cent of trisomy 2 1 pregnancies might be detected using a cut-off value of 0.5 MOM (multiples of the median) as the lower limit of normal for maternal serum AFP levels. Unfortunately, an 11 per cent false-positive rate accompanies this parameter. Alternatively, it has been observed that, using abnormal second-trimester maternal serum hCG values Addressee for correspondence: G. C. Wolf, M.D., Ph.D., University of New Mexico Hospital, 221 1 Lomas Blvd., N E Albuquerque, NM 87131-5287, U.S.A.
0197-3851/92/02009349$05.00 0 1992 by John Wiley & Sons, Ltd.
Received 15 April 1991 Accepted 8 July 1991
94
G . C. WOLF ET AL.
(either less than 0.25 MOM or greater than 2.5 MOM), 62 per cent (28 of 45) of aneuploid fetuses could be detected (Bogart et a/., 1987, 1989). Sensitivity could be increased to 83 per cent using hCG and ahCG in only Down’s syndrome pregnancies. Employing a limit of 2.0 MOM with hCG, Wald et a/. (1988) recorded a 55 per cent sensitivity and a 10 per cent false-positive rate among 77 trisomy 21 pregnancies. Muller and BouC (1990) reported a 64 per cent sensitivity when analysing hCG in 60 such pregnancies. Recent efforts have attempted to improve these rates by utilizing a combination of parameters, and have met with modest success. Thus, assessing maternal age, SPl, AFP, and hCG, 78 per cent of Down’s syndrome pregnancies were identified, with a false-positive rate of 3.4 per cent (Petrocik et al., 1990). Considering the acknowledged variations in hCG and realizing that evidence (Ozturk et a/., 1988)now suggests that the levels of both ahCG and bhCG are higher in amniotic fluid compared with maternal serum in the early second trimester, we felt it pertinent to investigate these parameters in chromosomally anomalous gestations. It was anticipated that the known differences in maternal serum values might be greatly exaggerated in the amniotic fluid compartment, and that perhaps one or more of these biochemical entities, or a ratio, might be of diagnostic value. A recent report regarding the ratio of peripheral levels of ahCG and bhCG in trisomy 18 pregnancies lends credence to our supposition (Ozturk et a/., 1990). MATERIALS AND METHODS Amniotic fluid specimens were obtained at the time of genetic amniocentesis at the University of New Mexico School of Medicine from 1982 to 1990. These samples were frozen at -20°C until retrieval. A total of 46 chromosomally abnormal fluids were recovered; simultaneously, at the time that these samples were selected, a control group of 92 specimens was obtained by choosing two vials, one chronologically preceding and one succeeding the intended sample; three control samples contained inadequate fluid to complete all assays. These specimens were than analyzed in duplicate for hCG, ahCG, and bhCG. Chromosome analyses had previously been performed by standard in situ trypsin G-banding techniques. The hCG was assayed (Hybritech Inc., San Diego, CA) utilizing a solid phase, two-site immunoradiometric system that recognizes both intact and free beta subunit and employs monoclonal antibody technology. The sensitivity was 5 mIU/ml and there was less than 0.1 per cent cross-reactivity with follicle stimulating hormone (FSH), luteinizing hormone (LH), and thyroid stimulating hormone (TSH). The assay was calibrated using the First International Reference Preparation (IRP) (75/537); intra- and inter-assay coefficients of variation were both less than 9 per cent. The bhCG was measured by a previously described assay using the monoclonal antibody IES, which detects only free subunit in the presence of the intact hCG molecule (Khazaeli et al., 1986). The cross-reactivity with hCG was 0.23 per cent while that with FSH, LH, and TSH was virtually zero. The ahCG was quantitated utilizing a commercially available radio-immunoassay system (Biomerica Inc., Newport Beach, CA) which also employs monoclonal antibody technology. The ahCG/ monoclonal antibody was precipitated with a second antibody (goat anti-mouse gamma-globulin); utilizing a radio-iodinated ahCG tracer, the amount of activity in
AMNIOTIC FLUID LEVELS OF HUMAN CHORIONIC GONADOTROPIN
95
Table 1. Summary of chromosomal abnormalities Group I Ia
Ib I1 IIa IIb 111
IIIa IIIb IV V VI
Karyotype
N
Normal
89
46,XY 46,XX Down’s syndrome 47,XY,+21 47,XX, 2 I Edward’s syndrome 47,XY,+ 18 47,Xx, I8
44 45 21 10 11 7 3 4 2 5 11
+
+
Turner’s syndrome Sex chrom. trisomy Miscellaneous
the precipitate was inversely proportional to the competing concentration in the sample. The assay was calibrated with the World Health Organization first IRP (hCG-alpha IRP No. 1; 75/569). Cross-reactivity with the intact glycoprotein hormones (FSH, LH, TSH, HCG) was less than 0.2 per cent; interference by the other free beta subunits was less than 0.1 per cent. The coefficients of variation were less than 12 per cent. The gestational age of each pregnancy was based on the results of the ultrasonographic evaluation performed just prior to the amniocentesis. The 46 chromosomally abnormal gestations are listed in Table I . There were three mosaic patterns included in these groups: one trisomy 21 contained 35 per cent 47,XY,+ 21 cells; one trisomy 18 exhibited 70 per cent 47,XX,+ 18 cells, and one sex trisomy displayed 92 per cent 47,XXX cells. Group V consisted of 47,XXX (n=2) and 47,XXY (n=3). The miscellaneous group (VI) contained the balanced translocations 46,XX,t( 14;22), 46,XX,t(3;6), 46,XX,t( 13;14), 45,XX,t( 14; 21), and 46,XY,t(1;22). The remaining anomalies in group VI were 46,X,del(X)(q22q26), 46,XY/46,XY,t(4;9), 47,XX,t mar, 46,XY/46,XY,9q - , 14q - ,- 15, t(9q, 14q,l5q), 47,XX, 13, and 47,XY, +9. Statistical analysis utilized non-parametric methods in the form of the KruskalWallis test after the Wilcoxon rank sums were obtained for each group. The Pearson correlation constants were generated by simple linear regression (SAS software program, SAS Institute, Cary, NC).
+
+
RESULTS Table 2 lists the descriptive statistics for the various groups and indicates the significant differences. With the exception of the difference in the gestational age in the subgroups of normal males and trisomy 18 males, the lengths of gestation among the groups are equivalent. Review of the individual medical records confirmed that maternal age was the major reason for amniocentesis in both controls (58/89) and study subjects (27/46); abnormal maternal serum AFP was the next most common reason, in 18/89 and 13/46 cases, respectively. Note that ahCG is significantly lower
96
G . C. WOLF ET AL.
Table 2. Statistical data (mean+SEM)
Group
I Ia Ib
I1 IIa IIb 111
IIIa IIIb IV V VI
Maternal age (years)
Gest. age (weeks)
35.0 f0.6 35.0 k 0.7 35.0 f0.8 38.8 f0.7 38.7f 1.1 38.9 'r 0.8 33.9 f2.9 3 5 7 k 1.4 32.5 f4.8 18.0 k0.7 36.8 k 1.O 34.3 1.6
16.2k0.1 16.0f0.2 16.4f0.2 16.3k0.4 16.9k0.6 15-7kO.3 15.8 f0.7 14.7f 1.0 16.6f0.8 18.5f1.8 14.9f0.7 15.9f0.9
ahCG (ng/ml)
bhCG (ng/ml)
589f42* 626f63 611 +56 561 k 6 3 569 f63 689 107 1286+182* 1293f397 324 k 86t 901 If:262 1636+201 2174+648? 1545f303* 409 k 82 1447f392 224f56 1620f438 548k86 480f14 995+346 1040f161* 567+207 1189f213 1232+276
hCG (mIU/ml)
ahCG/ bhCG
4758f423$ 4458+402 505 1 f737 7953k 1241$ 3244 _+ 6727 12234f132t 291 2 f6771 1551+357 3933+852 7133f1867 7907+1930 4823,248
0.94 1.09 0.83 0.99 2.79 0.75 3.78 6.46 2.96 0.48 1.83 0.97
*I < [II, 111, V]; p < 0.05. tIIa
12,93-10 1 ( 1992)
AMNIOTIC FLUID LEVELS OF HUMAN CHORIONIC GONADOTROPIN AND ITS ALPHA AND BETA SUBUNITS IN SECOND-TRIMESTER CHROMOSOMALLY ABNORMAL PREGNANCIES c. WOLF*, FRANCIS w. BYRN*,
THOMAS s. MCCONNELL~AND M. B. KHAZAELI~ *Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology and ?Department of Pathology, University of New Mexico Schoolof Medicine, New Mexico, U.S.A.;and SDeparrment of Medicine. University of Alabama at Birmingham, Alabama. V.S.A.
GORDON
SUMMARY Amniotic fluid from 135 pregnancies was assayed for human chorionic gonadotropin (hCG) and its free alpha (ahCG) and free beta (bhCG) subunits. Forty-six chromosomally abnormal pregnancies between 14 and 20 weeks’ gestation were matched with 89 chromosomally normal samples. Compared with controls, trisomy 21 pregnancies exhibited significantly elevated levels of all three peptides, whereas trisomy 18 gestations gave rise only to significant elevation of ahCG. Female fetuses in both the trisomy 21 and trisomy 18 pregnancies provided significantly elevated levels of hCG and bhCG compared to their male counterparts. On converting the values to multiples of the median, it was determined that 6 of 7 trisomy 18 samples had abnormally elevated alpha/beta ratios, as did 6 of 21 Down’s syndrome pregnancies. Further, 11 of 21 trisomy 21 gestations had abnormal amniotic fluid hCG levels. Using only ahCG, bhCG and their ratio, a 61 per cent sensitivity was found for these trisomies, with a 96 per cent specificity. KEY WORDS
Amniotic fluid hCG and subunits
INTRODUCTION In an effort to prenatally diagnose chromosomally anomalous pregnancies, various chemical and hormonaI parameters have recently come under scrutiny. These include alpha-fetoprotein (AFP) (Merkatz et al., 1984), unconjugated oestriol (UE3) (Canick et al., 1988), human chorionic gonadotropin (hCG) and its alpha subunit (Bogart et al., 1987; Petrocik et al., 1989) and beta subunit (Berkowitz et al., 1989), cancer antigen 125 (CA 125) (Check et al., 1990), Schwangershaft protein 1 (SPI) (Bartels and Lindemann, 1988), and combinations thereof (Wald et al., 1988; O’Brien et al., 1990; Petrocik et al., 1990). Evaluation of the association between low material serum AFP levels and fetal chromosome abnormalities suggests that approximately 30 per cent of trisomy 2 1 pregnancies might be detected using a cut-off value of 0.5 MOM (multiples of the median) as the lower limit of normal for maternal serum AFP levels. Unfortunately, an 11 per cent false-positive rate accompanies this parameter. Alternatively, it has been observed that, using abnormal second-trimester maternal serum hCG values Addressee for correspondence: G. C. Wolf, M.D., Ph.D., University of New Mexico Hospital, 221 1 Lomas Blvd., N E Albuquerque, NM 87131-5287, U.S.A.
0197-3851/92/02009349$05.00 0 1992 by John Wiley & Sons, Ltd.
Received 15 April 1991 Accepted 8 July 1991
94
G . C. WOLF ET AL.
(either less than 0.25 MOM or greater than 2.5 MOM), 62 per cent (28 of 45) of aneuploid fetuses could be detected (Bogart et a/., 1987, 1989). Sensitivity could be increased to 83 per cent using hCG and ahCG in only Down’s syndrome pregnancies. Employing a limit of 2.0 MOM with hCG, Wald et a/. (1988) recorded a 55 per cent sensitivity and a 10 per cent false-positive rate among 77 trisomy 21 pregnancies. Muller and BouC (1990) reported a 64 per cent sensitivity when analysing hCG in 60 such pregnancies. Recent efforts have attempted to improve these rates by utilizing a combination of parameters, and have met with modest success. Thus, assessing maternal age, SPl, AFP, and hCG, 78 per cent of Down’s syndrome pregnancies were identified, with a false-positive rate of 3.4 per cent (Petrocik et al., 1990). Considering the acknowledged variations in hCG and realizing that evidence (Ozturk et a/., 1988)now suggests that the levels of both ahCG and bhCG are higher in amniotic fluid compared with maternal serum in the early second trimester, we felt it pertinent to investigate these parameters in chromosomally anomalous gestations. It was anticipated that the known differences in maternal serum values might be greatly exaggerated in the amniotic fluid compartment, and that perhaps one or more of these biochemical entities, or a ratio, might be of diagnostic value. A recent report regarding the ratio of peripheral levels of ahCG and bhCG in trisomy 18 pregnancies lends credence to our supposition (Ozturk et a/., 1990). MATERIALS AND METHODS Amniotic fluid specimens were obtained at the time of genetic amniocentesis at the University of New Mexico School of Medicine from 1982 to 1990. These samples were frozen at -20°C until retrieval. A total of 46 chromosomally abnormal fluids were recovered; simultaneously, at the time that these samples were selected, a control group of 92 specimens was obtained by choosing two vials, one chronologically preceding and one succeeding the intended sample; three control samples contained inadequate fluid to complete all assays. These specimens were than analyzed in duplicate for hCG, ahCG, and bhCG. Chromosome analyses had previously been performed by standard in situ trypsin G-banding techniques. The hCG was assayed (Hybritech Inc., San Diego, CA) utilizing a solid phase, two-site immunoradiometric system that recognizes both intact and free beta subunit and employs monoclonal antibody technology. The sensitivity was 5 mIU/ml and there was less than 0.1 per cent cross-reactivity with follicle stimulating hormone (FSH), luteinizing hormone (LH), and thyroid stimulating hormone (TSH). The assay was calibrated using the First International Reference Preparation (IRP) (75/537); intra- and inter-assay coefficients of variation were both less than 9 per cent. The bhCG was measured by a previously described assay using the monoclonal antibody IES, which detects only free subunit in the presence of the intact hCG molecule (Khazaeli et al., 1986). The cross-reactivity with hCG was 0.23 per cent while that with FSH, LH, and TSH was virtually zero. The ahCG was quantitated utilizing a commercially available radio-immunoassay system (Biomerica Inc., Newport Beach, CA) which also employs monoclonal antibody technology. The ahCG/ monoclonal antibody was precipitated with a second antibody (goat anti-mouse gamma-globulin); utilizing a radio-iodinated ahCG tracer, the amount of activity in
AMNIOTIC FLUID LEVELS OF HUMAN CHORIONIC GONADOTROPIN
95
Table 1. Summary of chromosomal abnormalities Group I Ia
Ib I1 IIa IIb 111
IIIa IIIb IV V VI
Karyotype
N
Normal
89
46,XY 46,XX Down’s syndrome 47,XY,+21 47,XX, 2 I Edward’s syndrome 47,XY,+ 18 47,Xx, I8
44 45 21 10 11 7 3 4 2 5 11
+
+
Turner’s syndrome Sex chrom. trisomy Miscellaneous
the precipitate was inversely proportional to the competing concentration in the sample. The assay was calibrated with the World Health Organization first IRP (hCG-alpha IRP No. 1; 75/569). Cross-reactivity with the intact glycoprotein hormones (FSH, LH, TSH, HCG) was less than 0.2 per cent; interference by the other free beta subunits was less than 0.1 per cent. The coefficients of variation were less than 12 per cent. The gestational age of each pregnancy was based on the results of the ultrasonographic evaluation performed just prior to the amniocentesis. The 46 chromosomally abnormal gestations are listed in Table I . There were three mosaic patterns included in these groups: one trisomy 21 contained 35 per cent 47,XY,+ 21 cells; one trisomy 18 exhibited 70 per cent 47,XX,+ 18 cells, and one sex trisomy displayed 92 per cent 47,XXX cells. Group V consisted of 47,XXX (n=2) and 47,XXY (n=3). The miscellaneous group (VI) contained the balanced translocations 46,XX,t( 14;22), 46,XX,t(3;6), 46,XX,t( 13;14), 45,XX,t( 14; 21), and 46,XY,t(1;22). The remaining anomalies in group VI were 46,X,del(X)(q22q26), 46,XY/46,XY,t(4;9), 47,XX,t mar, 46,XY/46,XY,9q - , 14q - ,- 15, t(9q, 14q,l5q), 47,XX, 13, and 47,XY, +9. Statistical analysis utilized non-parametric methods in the form of the KruskalWallis test after the Wilcoxon rank sums were obtained for each group. The Pearson correlation constants were generated by simple linear regression (SAS software program, SAS Institute, Cary, NC).
+
+
RESULTS Table 2 lists the descriptive statistics for the various groups and indicates the significant differences. With the exception of the difference in the gestational age in the subgroups of normal males and trisomy 18 males, the lengths of gestation among the groups are equivalent. Review of the individual medical records confirmed that maternal age was the major reason for amniocentesis in both controls (58/89) and study subjects (27/46); abnormal maternal serum AFP was the next most common reason, in 18/89 and 13/46 cases, respectively. Note that ahCG is significantly lower
96
G . C. WOLF ET AL.
Table 2. Statistical data (mean+SEM)
Group
I Ia Ib
I1 IIa IIb 111
IIIa IIIb IV V VI
Maternal age (years)
Gest. age (weeks)
35.0 f0.6 35.0 k 0.7 35.0 f0.8 38.8 f0.7 38.7f 1.1 38.9 'r 0.8 33.9 f2.9 3 5 7 k 1.4 32.5 f4.8 18.0 k0.7 36.8 k 1.O 34.3 1.6
16.2k0.1 16.0f0.2 16.4f0.2 16.3k0.4 16.9k0.6 15-7kO.3 15.8 f0.7 14.7f 1.0 16.6f0.8 18.5f1.8 14.9f0.7 15.9f0.9
ahCG (ng/ml)
bhCG (ng/ml)
589f42* 626f63 611 +56 561 k 6 3 569 f63 689 107 1286+182* 1293f397 324 k 86t 901 If:262 1636+201 2174+648? 1545f303* 409 k 82 1447f392 224f56 1620f438 548k86 480f14 995+346 1040f161* 567+207 1189f213 1232+276
hCG (mIU/ml)
ahCG/ bhCG
4758f423$ 4458+402 505 1 f737 7953k 1241$ 3244 _+ 6727 12234f132t 291 2 f6771 1551+357 3933+852 7133f1867 7907+1930 4823,248
0.94 1.09 0.83 0.99 2.79 0.75 3.78 6.46 2.96 0.48 1.83 0.97
*I < [II, 111, V]; p < 0.05. tIIa
